217

Introduction

In his long career at the Bank of Canada, Chuck Freedman has focused notonly on monetary policy, but also on how to make the financial systemfunction more efficiently. This paper follows in the tradition set by Chuck byexamining policies to remedy conflicts of interest in the financial servicesindustry.

With the end of the stock market boom in 2000, financial markets have beenjolted by one corporate scandal after another. The cycle began with thespectacular bankruptcy of Enron Corporation in December 2001, oncevalued as the seventh largest corporation in the United States, and theindictment of Enron’s auditor, Arthur Andersen, one of the big fiveaccounting firms. Subsequently, there have been revelations of misleadingaccounting statements at numerous other corporations, includingWorldCom, Tyco Industries, and more recently, Ahold, which have added tothe doubts about the quality of accounting information in the corporatesector. Criminal cases have also been brought against investment banks forencouraging their stock analysts to hype stocks that they had serious doubtsabout and that turned out to be disastrous investments.

Policy Remedies for Conflicts ofInterest in the Financial System

Frederic S. Mishkin*

* I thank my discussant, Ron Parker, and other conference participants.

218 Mishkin

These scandals have received tremendous public attention, both becauseresulting bankruptcies have cost employees of these firms their jobs or theirpensions, and because of the subsequent stock market decline of over 40 percent (S&P 500) and 65 per cent (Nasdaq) from March 2000 to March 2003.At the root of these scandals may be conflicts of interest in which agentswho were supposed to provide the investing public with reliable informationhad incentives to hide the truth in order to further their own goals. What arethese conflicts of interest and how serious are they? Have they been thesource of recent financial market woes? What should be done about them?

This paper seeks to provide answers to these questions and is a summary ofa larger study that I have written with Andrew Crockett, Trevor Harris, andEugene White (Crockett et al. 2004). This paper outlines a framework foranswering these questions by first discussing the crucial role of informationin financial markets. This framework provides an understanding of whatconflicts of interest are and why we should care about them. The paper thenpresents a brief survey of the different types of conflicts of interest in thefinancial system. The paper concludes by developing a framework foranalyzing policies to remedy conflicts of interest and outlines specific policyrecommendations.

1 Information and Financial Markets

To understand why conflicts of interest are important, we need to step back abit and think about the function of financial markets in the economy. Well-functioning financial markets perform the essential economic function ofchannelling funds from individuals and firms who lack productiveinvestment opportunities to those who have such opportunities. By so doing,financial markets contribute to higher production and efficiency. Andreliable information is the key to this function.

A crucial impediment to the efficient functioning of the financial system isasymmetric information, a situation in which one party to a financial con-tract has much less accurate information than the other party. For example,managers of corporations usually have much better information about thepotential returns and risks associated with the investment projects they planto undertake than do potential purchasers of the corporation’s stock.Asymmetric information leads to two basic problems: adverse selection andmoral hazard.

Adverse selection is an asymmetric information problem that transpiresbefore the transaction occurs, when parties who are the most likely toproduce an undesirable (adverse) outcome for a financial contract are mostlikely to try to enter the contract and thus beselected. For example,

Policy Remedies for Conflicts of Interest in the Financial System 219

managers who want to siphon funds for their personal use are likely to be themost eager to enlarge their company by raising funds. Since adverseselection makes it more likely that investments in firms will turn out badly,investors may decide not to invest even if there are attractive investments inthe marketplace. This outcome is a feature of the classic “lemons problem”analysis first described by Akerlof (1970). Clearly, minimizing the adverse-selection problem so that capital flows to productive use requires that inves-tors have the information to screen out good from bad investments.

Moral hazard occurs after the transaction takes place because the provider offunds is subjected to thehazardthat the receiver of funds has incentives toengage in activities that are undesirable from the lender’s point of view (i.e.,activities that make it less likely that the investment will be a good one).Moral hazard occurs because the receiver of funds has incentives tomisallocate funds for personal use or to undertake investment in unprofitableprojects that increase personal power or stature. As a result, many investorswill decide that they would rather not provide firms with funds, so thatinvestment will be at suboptimal levels. To minimize the moral-hazardproblem, investors must have information so that they can monitor theactivities of managers to ensure that they use the funds to maximize thevalue of the firm.

As the discussion of asymmetric information problems of adverse selectionand moral hazard illustrates, the provision of reliable information is crucialto the ability of financial markets to perform their essential function ofchannelling funds to those with productive investment opportunities. Forinvestors to be willing to provide funds for investment projects, they must beable to screen out good from bad credit risks in order to avoid the adverse-selection problem, and they also need to monitor those to whom theyprovide funds in order to minimize the moral-hazard problem. But how isthe information that investors require to be provided?

2 The Role of Financial Institutions in Financial Markets

An obvious answer to the question above is that private investors couldcollect the necessary information themselves to enable them to screen andmonitor their investments. There are two barriers to their doing so, however.First is the free-rider problem.

The free-rider problem occurs when people who do not spend resources oncollecting information can still take advantage of (free ride off) theinformation that others have collected. The free-rider problem is particularlyimportant in securities markets. If well-informed investors are able to buy inadvance of others on the basis of their superior research, they can benefit

220 Mishkin

from their superior information. But, if other investors who have not paid forthis information obtain it quickly enough, they may be able to capture someof the value. If enough free-riding investors can do this, investors who haveacquired information will no longer be able to earn the entire increase in thevalue of the security arising from this additional information. The weakenedability of private investors to profit from producing information will meanthat less information is produced in securities markets, so that the adverse-selection problem, in which overvalued securities are those most oftenoffered for sale, is more likely to impede a well-functioning securitiesmarket.

Possibly even more important, the free-rider problem makes it less likelythat there will be sufficient monitoring to reduce incentives to commit moralhazard. By monitoring borrowers’ activities to determine whether they arecomplying with restrictive covenants and enforcing the covenants if they arenot, lenders can prevent borrowers from taking on risk at their expense.Similarly, the monitoring of managers can help ensure that they do not divertfunds for their personal use or make expenditures that bring them prestige orperquisites rather than raise shareholder value. However, because moni-toring is costly, the free-rider problem discourages this kind of activity insecurities markets. If some investors know that other securities holders aremonitoring and enforcing restrictive covenants, they can free ride on themonitoring and enforcement of other securities holders. Once these othersecurities holders realize that they can do the same thing, they may also stoptheir monitoring and enforcement activities, with the result that not enoughresources are devoted to monitoring and enforcement. The outcome is thatmoral hazard is likely to be a severe problem in financial markets.

Financial institutions can help mitigate the free-rider problem by acquiringfunds from the public and using them to purchase and hold assets in adiversified portfolio based on the specialized information they collect. Asfinancial intermediaries, they can act as delegated monitors (Leland and Pyle1977). They are not as subject to the free-rider problem and profit from theinformation they produce because they can make investments such as bankloans that are often not traded. Because the investments are not marketed,other investors cannot buy them. As a result, investors are less able to freeride off financial intermediaries and bid up the prices of the securities, whichwould prevent the intermediary from profiting from its information pro-duction activities. Similarly, it is hard to free ride off these financial inter-mediaries’ monitoring activities when they make bank loans. Financialinstitutions making private investments thus receive the benefits ofmonitoring and are better equipped to prevent moral hazard on the part ofborrowers or managers.

Policy Remedies for Conflicts of Interest in the Financial System 221

While this strategy works for non-depository intermediaries if their share-holders participate in the information discovery or are given signals by themanagers, depository intermediaries would be subject to the same challengeas businesses in signalling the value of the portfolio of assets in which theyhave invested. Rather than signal by their own substantial holding ofdeposits, the solution for depository intermediaries is the issue of demand-able deposits. Deposits that are quickly redeemable enable depositors todiscipline managers if they believe that risk has increased by withdrawingtheir funds (Calomiris and Kahn 1991).

A second barrier to private production of information is that investors maynot be able to sufficiently diversify or operate on a sufficient scale so thatinformation production is too costly. Financial institutions are able to reacha sufficient size so that they can diversify and reduce average screening andmonitoring costs (Diamond 1984; Ramakrishnan and Thakor 1984). How-ever, a financial institution must persuade the primary investors that it is ade-quately monitoring the business it is funding. To do this, it must conductinternal monitoring of its employees so that they engage in the appropriatelevel of screening and monitoring of investments.

In the literature described thus far, financial institutions are treated asfocusing on only one type of informational asymmetry. Indeed, one couldrationalize many different types of financial institutions on the grounds thateach type addresses a different informational asymmetry. However, theinformation that any one institution possesses may be useful beyond theprovision of one narrow type of service. For instance, banks, owing to theirlong-term customer relationships, obtain reusable private information aboutfirms’ resources, cash flows, and other characteristics. For individual cus-tomers, they gather information, often confidential, beyond what is publiclyavailable, and this information is obtained by the provision of services overtime. The closeness of a relationship over time may induce the customer toreveal more confidential information and thereby gain some advantage withthe financial firm (Boot 2000).

Financial institutions gain a cost advantage in the production of informationbecause they develop special skills to interpret signals and exploit cross-sectional information across customers. Furthermore, the reusability ofinformation gives them another advantage as the initial informationproducer specializing in its production and distribution (Chan, Greenbaum,and Thakor 1986; Greenbaum and Thakor 1995). Thus, not only are theylower-cost producers of information for one type of financial service, butthey can also be lower-cost producers of information for multiple financialservices, which become complementary activities. It is also usually assumedthat institutions that combine several financial services have advantages over

222 Mishkin

specialized ones. By providing a broader set of financial products, aninstitution can develop wider and longer-term relationships to firms that maybe the source of further economies of scope (Santos 1998). A financialinstitution may learn more about a firm by the provision of more types offinancial instruments from which it can collect more varied information andwhich may give it more monitoring and disciplinary power.

3 What Are Conflicts of Interest?

While the presence of the synergies or economy of scope described abovemay offer substantial benefits, they also create potential costs in the form ofconflicts of interest. These conflicts exist “whenever one is serving two ormore interests and can put one person in a better position at the expense ofanother” (Edwards 1979). Because conflicts of interest are present in almostall aspects of our lives, we need to be more precise about those of concernhere. Given the crucial role of information in financial markets, this paperuses the following definition for conflicts of interest:

Conflicts of interest arise when a financial service provider, oran agent within such a service provider, has multiple intereststhat create incentives to act in such a way as to misuse orconceal information needed for the effective functioning offinancial markets.

Conflicts of interest may occur within specialized financial institutions.However, conflicts of interest stand out most sharply when an institutionprovides multiple financial services, thereby creating an opportunity forexploiting the synergies or economies of scope by inappropriately divertingsome of their benefits. Combinations of services that bring together anygroup of depository intermediaries, non-depository intermediaries, orbrokers or allow any of these to directly invest in business have attracted thegreatest criticism for putative conflicts of interest.

4 Why Conflicts of Interest Are Important

We care about these conflicts of interest because if they sufficiently reducethe amount of information in financial markets, they increase asymmetricinformation and prevent financial markets from channelling funds to thosewith productive investment opportunities. There are clearly broaderdefinitions of conflicts of interest than the one stated above, and many ofthese conflicts of interest are exploited to the detriment of individuals andthe economy. However, this paper restricts itself to a narrower view, becauseconflicts of interest require public policy intervention only if they lead toless efficient financial markets.

Policy Remedies for Conflicts of Interest in the Financial System 223

5 Types of Conflicts of Interest

There are four areas of financial service activities that have the greatest po-tential for conflicts of interest that reduce information in financial markets.They are described below.

5.1 Underwriting and research in investment banking

The information synergies from underwriting, research, and market makingprovide a rationale for combining these distinct financial services. This com-bination of activities leads to conflicts of interest, however. The conflict ofinterest that raises the greatest concern occurs between underwriting andbrokerage, where investment banks are serving two client groups, issuingfirms and investors. Issuers benefit from optimistic research, while investorsdesire unbiased research. If the incentives for these two activities are notappropriately aligned, there will be a temptation for employees on one sideof the firm to distort information to the advantage of their clients and theprofit of their department. When the potential revenues from underwritinggreatly exceed brokerage commissions, there will be a strong incentive tofavour issuers over investors or risk losing the former to competitors. As aresult, analysts in investment banks might distort their research to pleaseissuers, and the information they produce on securities will not be asreliable, thereby diminishing the efficiency of securities markets.

5.2 Auditing and consulting in accounting firms

The traditional role of an auditor has been to act as an efficient monitor ofthe quality of information produced by firms so as to reduce the inevitableinformation asymmetry between the firm’s managers and stakeholders,especially its suppliers of capital. In auditing, threats to truthful reportingarise from several potential conflicts of interest that can lead to biasedoutcomes. The conflict that has received the most attention lately occurswhen an accounting firm provides auditing as well as non-audit consultingservices—tax advice, accounting, or management information systems, andstrategic advice, commonly referred to as management advisory services.These multiple services enjoy economies of scale and scope, but create twopotential sources of conflict of interest. The most commonly discussedconflict is the potential to pressure auditors to bias their judgments andopinions to limit any loss of fees in the “other” services. The second moresubtle conflict is that auditors often evaluate systems or structuring (tax andfinancial) advice that were put in place by their non-audit counterpartswithin the firm. Both conflicts may lead to biased audits, with the result that

224 Mishkin

less information is available in financial markets, which will make it harderfor them to efficiently allocate capital.

5.3 Credit assessment and consulting in rating agencies

Ratings are widely used by investors as a guide to the creditworthiness ofthe issuers of debt, and in financial covenants. As such, they play a majorrole in the pricing of debt securities and in the regulatory process. Conflictsof interest can arise from the fact that there are multiple users of ratings;and, at least in the short term, their interests can diverge. The investor andregulators are interested in a well-researched, impartial assessment of creditquality; the issuer in a favourable rating. Because issuers pay to have theirsecurities rated, there is a fear that credit agencies may bias their ratingsupwards in order to attract more business. A more serious concern is thatrating agencies have begun to provide ancillary consulting services in recentyears. Rating agencies are increasingly asked to advise on the structuring ofdebt issues, usually to help secure a favourable rating. In this case, the creditrating agency would be in the position of “auditing its own work,” raisingconflicts of interest similar to those in accounting firms when they provideboth auditing and consulting services. Furthermore, providing consultingservices creates additional incentives for the rating agencies to deliver morefavourable ratings in order to further their consulting business. The possiblereduction in the quality of credit assessment by rating agencies could thenincrease asymmetric information in financial markets, thereby reducing theirability to allocate credit.

5.4 Universal banking

Although commercial banks, investment banks, and insurance companiesoriginally arose as distinct financial institutions, there were economies ofscope that could be attained by their combination, thus leading to thedevelopment of universal banking in which all of these activities arecombined in one organization. Yet, given that activities within a universalbank serve multiple clients, there are many potential conflicts of interest. Ifthe potential revenues from one department surge, there will be an incentivefor employees in that department to distort information to the advantage oftheir clients and the profit of their department. For example, issuers servedby the underwriting department will benefit from aggressive sales tocustomers of the bank, while these customers are hoping to receive unbiasedinvestment advice. A bank manager may push the affiliate’s products to thedisadvantage of the customer or limit losses from a poor public offering byplacing them in the bank’s managed trust accounts. A bank with a loan to afirm whose credit or bankruptcy risk has increased, has private knowledge

Policy Remedies for Conflicts of Interest in the Financial System 225

that may encourage it to use the bank’s underwriting department to sellbonds to the unsuspecting public, thereby paying off the loan and earning afee. A bank may make loans on overly favourable terms in order to obtainfees from activities like underwriting securities. To sell its insuranceproducts, a bank may attempt to influence or coerce a borrowing or investingcustomer. All of these conflicts of interest may lead to a decrease in accurateinformation production by the universal bank, thereby hindering its ability topromote efficient credit allocation.

6 When Are Conflicts of Interest a Serious Problem?

The analysis of conflicts of interest here begins with the observation thatthey present their main problem for the financial system when they lead to adecrease in information flows that makes it more difficult for the system tosolve adverse-selection and moral-hazard problems that can slow the flow ofcredit to parties with productive investment opportunities. Even though aconflict of interest exists, it does not necessarily reduce the flow ofinformation, because the incentives to exploit the conflict of interest may notbe very high. Exploitation of a conflict of interest that is visible to themarket will typically result in a decrease in the reputation of the financialfirm involved. Given the importance of maintaining and enhancingreputation, exploiting the conflict of interest would then decrease the futureprofitability of the firm, because it will have greater difficulty selling itsservices in the future, thus creating incentives for the firm to preventexploitation of the conflict of interest. The evidence surveyed in Crockett etal. (2003) suggests that these incentives do work to constrain conflicts ofinterest in the long run, but the extent to which they are effective in the shortrun depends on factors such as transparency and incentive structures withinfirms.

One example occurs in credit rating agencies. At first glance, the fact thatthey are paid by the firms issuing securities to produce ratings for thesesecurities looks like a serious conflict of interest. Rating agencies wouldappear to have incentives to gain business by providing firms issuingsecurities with higher credit ratings than they deserve, making it easier forthem to sell these securities at higher prices. However, there is little evidencethat rating agencies engage in this conflict of interest: considerable research(surveyed in Bank for International Settlements 2000) has shown that,despite prominent counter-examples, such as Enron, there is a reasonablyclose correlation between ratings and default probabilities. The conflict ofinterest does not appear to be exploited because giving higher credit ratingsto firms that pay for the ratings would result in decreased credibility of theratings, thus making them less valuable to the market. The market is

226 Mishkin

eventually able to assess the quality of biased ratings down the road becauseit can observe poorer performance by individual securities. Furthermore,rating agencies themselves provide evidence on the relationship betweentheir ratings and subsequent default history (Brand and Bahar 1999; Keenan1999). The resulting loss of trust in the information provided by the ratingagency if this conflict were to be exploited would lead to a costly decline inits reputation, thus providing incentives not to exploit this conflict ofinterest.

Similarly, the apparent conflicts of interest when commercial banks under-wrote securities before the Glass-Steagall Act do not appear to have beengenerally exploited. When a commercial bank underwrites securities, thebank may have an incentive to market the securities of financially troubledfirms to the public because the firms will then be able to pay back the loansthey owe to the bank, while the bank earns fees from the underwritingservices. Evidence for the 1920s suggests that this conflict of interest causedmarkets to find securities underwritten by bond departments within acommercial bank to be less attractive than securities underwritten inseparate affiliates where the conflicts of interest were more transparent(Kroszner and Rajan 1994). In order to maintain the bank’s reputation,commercial banks shifted their underwriting to separate affiliates over time,with the result that securities underwritten by banks were valued as highly asthose underwritten by independent investment banks (Ang and Richardson1994; Puri 1996; Kroszner and Rajan 1997). When affiliates were unable tocertify the absence of conflicts, they focused on more senior securitieswhere there was less of an information asymmetry and conflicts were lesssevere. Again, the market provided incentives to control potential conflictsof interest. It is important to note, however, that the market solution was notimmediate, but took some time to develop.

The responsiveness of the market can also be seen in the apparent conflict ofinterest for investment banks when underwriters who have incentives tofavour issuers over investors pressure research analysts to provide morefavourable assessments of issuers’ securities. It has been observed that leadunderwriters make more buy recommendations for initial public offerings(IPOs) than other firms’ analysts for the same securities, yet the stock pricesof firms recommended by lead underwriting investment banks declinedduring the Securities and Exchange Commission’s (SEC’s) 25-day quietperiod while other banks’ choices rose (Michaely and Womack 1999). Overa two-year period, the performance of other analysts’ recommended issueswas 50 per cent better than the performance of underwriters’ recom-mendations. Hence, the market appears to recognize the difference in thequality of information when there is a potential conflict of interest.

Policy Remedies for Conflicts of Interest in the Financial System 227

There are fewer empirical studies in auditing, but even this limited evidencesuggests that the market perceives and adjusts for potential conflicts ofinterest. There is evidence that clients who are concerned that conflicts ofinterest from the joint provision of auditing and management advisoryservices will reduce the value their audit opinions limit non-audit purchasesfrom incumbent auditors (Parkash and Venable 1993).

These examples do not indicate that the market can always contain theincentives to exploit conflicts of interest. For the market to preventexploitation of conflicts, it needs to have information on whether thisexploitation might be occurring. In some cases, divulging information thatwould reveal whether a conflict of interest is being exploited may not beavailable to the market. In other cases, providing such information wouldreveal proprietary information that would help a financial firm’s com-petitors, thus reducing the incentives to reveal this information.

As brought out in the recent scandals, what are particularly worrisome areconflicts of interest whose exploitation leads to large gains for somemembers of the financial firm even if it reduces the value of the entire firm.Compensation mechanisms inside a firm, if inappropriately designed, maylead to conflicts of interest that not only reduce information flows to creditmarkets but end up destroying the firm. In other words, damaging conflictsof interest are likely to arise from poor management. Indeed, the story of thedemise of Arthur Andersen illustrates how the compensation arrangementseven for one line of business, such as auditing, can create severe conflicts ofinterest, in this case because partners in regional offices had incentives toplease their largest clients even if this was detrimental to the overall firm.The conflict of interest problem can become even more severe when severallines of business are combined and the returns from a particular activity—underwriting, consulting—are very high and expected to be brief, so that acompensation scheme that worked reasonably well at one time might be-come very badly aligned.

The extraordinary surge in the stock market created huge temporaryrewards, permitting well-positioned analysts, underwriters, or audit firmpartners to take advantage of the conflicts before incentives could berealigned. The reason these conflicts of interest are so dangerous is that theyare not readily visible to the market and may not even be visible to the topmanagement of the firm. In the most severe cases, opportunistic individualswere able to capture the firm’s reputational rents. The exploitation of theseconflicts of interest clearly damaged the reputation of such investment banksas Merrill Lynch, Solomon-Smith-Barney of Citigroup, and Credit SuisseFirst Boston, and perhaps the credibility of analysts in general. Audit firmshave lost much of their non-audit business, while Arthur Andersen wasdestroyed.

228 Mishkin

7 A Framework for EvaluatingPolicies to Remedy Conflicts of Interest

The information view of conflicts of interest proposed here also provides aframework for evaluating whether they require public policy actions toeliminate or reduce them. Some combination of financial service activitiesmay result in incentives for agents to conceal information, but they may alsoresult in synergies that make it easier to produce information. Thus, pre-venting the combination of activities to eliminate the conflicts of interestmay actually make financial markets less efficient. This reasoning suggeststhat there are two propositions that are critical to evaluating what should bedone.

• The fact that a conflict of interest exists does not mean that it will haveserious adverse consequences.Although a conflict of interest exists, theincentives to exploit it may not be very high. Exploitation of a conflict ofinterest that is visible to the market will typically result in a decrease inthe reputation of the financial firm involved. Given the importance ofmaintaining and enhancing reputation, exploiting the conflict of interestwould then decrease the future profitability of the firm because it willhave greater difficulty selling its services, thus creating incentives for thefirm to avoid taking advantage of the conflict of interest. Hence, themarketplace may be able to control conflicts of interest because of thehigh value of a firm’s reputation. When evaluating the need for remedies,this proposition raises the issue of whether the market has the infor-mation and incentives to control conflicts of interest.

• Even if incentives to exploit conflicts of interest remain strong, elimi-nating the conflict of interest may be harmful if doing so destroyseconomies of scope, thereby reducing information flows.In evaluatingpossible remedies, therefore, we also need to examine the issue ofwhether imposing the remedy will do more harm than good by reducingthe flow of information.

In considering remedies for specific conflicts, it is worth discussing fivegeneric approaches. These approaches are listed in the order of their intru-siveness, from least to most intrusive.

(i) Leave it to the market.This approach has a powerful appeal to manyeconomists, and may be a sufficient response in many cases. Marketforces can work through two mechanisms. First, they can penalize theservice provider if they exploit conflicts of interest. For example, apenalty may be imposed by the market in the form of higher fundingcosts or lower demand for its services, in varying degrees, even to thepoint of forcing the demise of the firm. Second, market forces can

Policy Remedies for Conflicts of Interest in the Financial System 229

promote new institutional means to contain conflicts of interest, forexample, by generating a demand for information from non-conflictedorganizations. This is exactly what happened when security affiliatestook pre-eminence over in-house bond departments in universal banksin the United States in the 1920s.

The advantages of market-driven solutions include the fact that they can hitwhere it hurts most—through pecuniary penalties. Moreover, they may helpavoid the risk of overreaction. It can be difficult to resist the temptation toadopt non-market solutions to appease public opinion that may reduceinformation production in financial markets. On the other hand, market-based solutions may not always work if the market cannot obtain sufficientinformation to appropriately punish financial firms that are exploitingconflicts of interest. Memories may be short in financial markets; once atriggering event has faded from memory, conflicts may creep back in unlessreforms have been “hard-wired.”

(ii) Regulate for transparency.A competitive market structure does notalways adequately reduce information asymmetries. The gathering ofinformation is costly, and any individual economic agent will gatherinformation only if the private benefit outweighs the cost. When theinformation becomes available to the market, the free-rider problemmay become serious. Information has the attribute of a public good,which will be undersupplied in the absence of public intervention. Tosome extent, mandatory information disclosure can alleviate infor-mation asymmetries and is a key element of regulation of the financialsystem.

Mandatory disclosure of information that reveals whether a conflict ofinterest exists may help the market to discipline financial firms that engagein conflicts of interest. In addition, if a financial institution is required toprovide information about potential conflicts of interest, the user of theinstitution’s information services may be able to judge how much weight toplace on the information this institution supplies.

On the other hand, mandatory disclosure could create problems if it revealsso much proprietary information that the financial institution is unable toprofitably engage in the information production business. The result couldthen be less information production rather than more. Also, mandatorydisclosure may not work if financial firms are able to avoid the regulationand continue to hide relevant information about potential conflicts ofinterest. The free-rider problem might also result in insufficient monitoringof conflicts of interest, because the benefits of monitoring and constrainingthese conflicts accrue only partially to the monitors.

230 Mishkin

(iii) Supervisory oversight.If mandatory disclosure does not work becausefirms are still able to hide relevant information, because the free-riderproblem is severe or because mandatory disclosure would revealproprietary information, supervisory oversight can come to the rescueand contain conflicts of interest. Supervisors can observe proprietaryinformation about conflicts of interest without revealing it to afinancial firm’s competitors so that the firm can continue to profitablyengage in information production activities. Supplied with thisinformation, the supervisor can take actions to prevent financial firmsfrom exploiting conflicts of interest. As part of this supervisoryoversight, standards of practice can be developed, either by thesupervisor, or by the firms engaged in a specific information-production activity. Enforcement of these standards would then be inthe hands of the supervisor.

Supervisory oversight of this type is very common in the banking industry.In recent years, bank supervisors have increased their focus on riskmanagement. They examine a bank’s risk-management procedures to ensurethat the appropriate internal controls on risk taking are in place. In a similarfashion, supervisors can examine the internal procedures and controls torestrict conflicts of interest. When they find weak internal controls, they canrequire the financial institution to modify them so that incentives to engagein conflicts of interest are eliminated.

Although supervisory oversight has been successful in improving internalcontrols in financial firms in recent years, if the incentives to engage inconflicts of interest are sufficiently strong, financial institutions may be ableto hide conflicts of interest from the supervisors. Furthermore, as recentbanking crises illustrate, supervisors have sometimes engaged in regulatoryforbearance in which they do not sufficiently enforce penalties on financialfirms engaged in undesirable behaviour. There is always the issue of whethersupervisors can be sufficiently insulated from short-term political pressuresto let financial institutions off the hook (i.e., avoid regulatory capture) andcan be made sufficiently accountable to prevent conflicts of interest fromgetting out of hand. On the other side, supervisors could become over-bearing and interfere with the efficient function of financial firms in order toavoid having a scandal occur on their watch.

(iv) Separation of functions.Where the market cannot obtain sufficientinformation to constrain conflicts of interest because there is nosatisfactory way of inducing information disclosure by marketdiscipline or supervisory oversight, the incentives to exploit conflictsof interest may be reduced or eliminated by regulations enforcingseparation of functions. There are several degrees of separation. First,there is separation of activities into different in-house departmentswith firewalls between them. The second degree is to conduct different

Policy Remedies for Conflicts of Interest in the Financial System 231

activities in separately capitalized affiliates. The third is prohibition ofthe combination of activities in any organizational form.

Separation of functions has the goal of ensuring that “agents” are not placedin the position of responding to multiple “principals” so that conflicts ofinterest are reduced. Moving from less to more stringent separation offunctions, conflicts of interest are reduced. However, more stringentseparation of functions reduces synergies of information collection, therebypreventing financial firms from taking advantage of economies of scope ininformation production. Deciding on the appropriate amount of separationthus involves a trade-off between the benefits of reducing conflicts ofinterest and the cost of reducing economies of scope in producinginformation.

(v) Socialization of information production.The most radical response toconflicts of interest generated by asymmetric information is tosocialize the provision or the funding source of the relevantinformation. For example, much macroeconomic information isprovided by publicly funded agencies, recognizing the argument thatthis particular public good is likely to be undersupplied if left toprivate provision. It is conceivable that other information-providingfunctions, for example, credit ratings and auditing, could also bepublicly supplied. Alternatively, if the information-generating servicesare left to the private sector, they could be funded by public sources orby a publicly mandated levy to help ensure that information productionis not tainted by obligations to fee-paying entities with specialinterests.

The problem with this approach, of course, is that a government agency orpublicly funded entity may not have the same incentives as private financialinstitutions to produce high-quality information. Forcing informationproduction to be conducted by a government or quasi-government entity,although it may reduce conflicts of interest, may result in the reduction ofinformation flow to financial markets. Furthermore, there is a compensationproblem in government agencies, because they may not be able to paymarket wages to attract the best people. This problem may be even moresevere if there are economies of scope: for example, analysts in an invest-ment banking firm are likely to receive additional compensation when theirresearch has multiple uses. A government agency interested in only one useof research may not provide a level of compensation adequate to producehigh-quality information.

232 Mishkin

8 Evaluating Remedies

The above framework suggests that in evaluating specific conflicts ofinterest, it is important to ask two questions:

• Do markets have the information and incentives to control conflicts ofinterest? As we have seen, the marketplace may be able to controlconflicts of interest because there is a high value to the reputation offinancial firms.

• Even if the incentives to exploit a conflict of interest are strong, would apolicy that eliminates the conflict of interest destroy economies of scope,thereby reducing information flows?

If the answer to either question is yes, the case for a policy to remedy aparticular conflict of interest is substantially weakened. Putting the remedyinto practice would then likely reduce the overall information in the market-place, thus doing more harm than good.

In designing appropriate remedies, it is important to remember that conflictsof interest did not create the boom or bubble in the stock market. Rather, theconflicts were opportunities to exploit the very rapid rise of stock prices incertain sectors. Conflicts may be largely eliminated by a completeseparation or segregation of each type of financial activity, but that wouldclearly entail a huge cost by drastically reducing the economies of scope.Stock market booms are infrequent events. The only cases that parallel theevents of the late 1990s are the late 1920s and possibly 1986–87. To imposesegregation remedies on the financial industry to prevent the exploitation ofconflicts in the rare spectacular bull markets will result in excessively highcosts. The imposition of the Glass-Steagall Act’s separation of commercialand investment banking after the boom of the 1920s is a clear example of anexcessive response that imposed large and unnecessary long-term costs onthe financial industry. Exploitation of conflicts of interest examined herewas never uniform across each industry, and litigation may be theappropriate response to discipline specific firms and individuals as part of anoverall market solution. However, legal liabilities and penalties need to becarefully designed, as witnessed by the behaviour of audit firms seeking toavoid the extremely high litigation risk from class-action lawsuits.

In evaluating remedies it is also important to remember that there are manytypes of agents in the financial system who provide information to themarket, ranging in degree from those with the least access to proprietaryinformation to those with the most. Analysts have the least access, andrating agencies have more. Auditors probably have the most privilegedaccess along with government regulators charged with supervisoryoversight. This scale of access to proprietary information should reflect the

Policy Remedies for Conflicts of Interest in the Financial System 233

ability of agents to discover the true financial condition and performance ofthe firms that they observe. Their ability to discover this information willalso be determined by their compensation and other incentives provided tothem. Although these agents provide some overlapping information, one isnot a substitute for another. This lack of substitution is not solely becausethey provide different types of information or signals to the public. Theseagents are all subject to various pressures and conflicts of interest that maydiminish their ability to perform their task of discovery. Analysts may bewell compensated and have substantial research resources at their disposal,but they may be too favourable to the firms for which their bank is leadunderwriter and they have the least access to proprietary information.Ratings agencies are largely insulated from conflicts of interest and havebetter access to proprietary information; but enjoying an oligopoly, theirresearch effort may be reduced. Auditors enjoy superior access toproprietary information and operate in a competitive industry, but the valueof their opinions may be reduced by conflicts between audit and non-auditactivities and a litigation-risk-induced focus on rules rather than principles.Finally, regulators/supervisors may have the best access to proprietaryinformation, yet their capacity to monitor is limited by the resources theyhave been allocated and political pressures for forbearance. To ensure thatthe capital markets are adequately served, it is necessary to have multipleagents working to reduce the information asymmetries. One may becomeless useful at one point, but maintaining the quality of information deliveredby these different agents engaged in overlapping work is more likely toprovide sufficient monitoring. Appropriate remedies should increase theeffectiveness of all four types of agents.

The review of the evidence on conflicts of interest in this paper suggests thatthe market is often able to constrain conflicts of interest to a considerabledegree, even though at first glance they seem to be severe and there is alearning process. Furthermore, it is dangerous to prevent exploitation ofsynergies in information production because this could substantially reducethe amount of information available in financial markets, thereby reducingthe efficiency of these markets in channelling funds to those with productiveinvestment opportunities. I may be showing my bias as an economist, butwhen it can be made to work, the market is the most effective and desirableway of disciplining conflicts. So the first focus of solutions to remedy con-flicts of interest should be on strengthening market discipline. Only whenone is convinced that market discipline cannot constrain serious conflicts ofinterest that reduce information flows, should they recommend non-marketsolutions. It is also important to note that market solutions work in the longrun; non-market solutions work in the short run, but they can hinder orprevent the emergence of more efficient market solutions in the long run.

234 Mishkin

9 Policy Recommendations

Using the information-oriented framework developed above leads to thefollowing recommendations on remedies for conflicts of interest in thefinancial services industry.

(i) Increase disclosure for investment analysts, credit rating analysts, andauditors to reveal any interests they have in the firms they analyze.Disclosure plays an important role, enabling markets to acquireinformation that can be used to punish financial firms that exploitconflicts of interest. Provision of this information makes it more likelythat financial services firms will develop internal rules to ensure thatconflicts of interest are minimized so that their reputation remains high,thus enabling them to continue to profitably engage in the information-production activities. Recent efforts by the SEC and other governmentagencies to increase disclosure of conflicts of interest are moves in theright direction.

(ii) Improve corporate governance.Remedies for controlling conflicts ofinterest cannot be effective in a vacuum. Without good corporategovernance, markets are unlikely to work well and so the remediesdiscussed here would be unlikely to solve conflict of interest problems.Improving corporate governance is a huge topic that is well beyond thescope of this paper. However, there is one particular area of corporategovernance that is critical to the quality of information in the financialsystem. Auditors need to be hired by, compensated by, and report toaudit committees whose responsibility is to represent stakeholdersother than management, as provided for in the Sarbanes-Oxley Act.Proper implementation of this reform is an important function for thePublic Company Accounting Oversight Board (PCAOB).

(iii) Establish codes of conduct developed by industry participants in co-operation with supervisors.Given their experience, financial serviceproviders in the private sector are capable of designing effectiveinternal controls and codes of conduct. But government supervisorscan help because they can monitor internal controls at many firms andobserve what is best practice. It is important that these codes bedynamic. The marketplace in financial services is in a continuous stateof flux, and best practice to control conflicts of interest will of neces-sity change over time.

(iv) Increase supervisory oversight.Mandatory disclosure may not alwaysbe sufficient to enable the market to constrain conflicts of interest,especially as it may be necessary to limit disclosure of proprietaryinformation. There is thus a strong role for supervisory oversight. Ithas an important role in containing conflicts of interest, because manyof the most damaging conflicts of interest arise from agency problems

Policy Remedies for Conflicts of Interest in the Financial System 235

within firms, the result of poorly designed internal compensationmechanisms that are difficult for markets to observe.

Banking supervisors already have powers to supervise universal banks andto monitor their internal control procedures to make sure that they do nottake on excessive risk. Bank supervision has been expanded in recent yearsto focus onoperational risk, and conflicts of interest can easily be viewed asa form of operational failure. Thus, the focus of bank supervisors onuniversal banks’ internal controls and compensation mechanisms withregard to conflicts of interest is a natural development. Controlling conflictsof interest in universal banks also has a growing importance for preservingthe safety and soundness of banks (and so is important from a prudentialperspective) because banks may lend on favourable terms in order to obtainfees from other activities, like underwriting securities. Just as bank super-vision has become more oriented to focus on risk management in recentyears, it needs to increase its focus on control of conflicts of interest.

The SEC and its equivalents in other countries have a clear interest in theactivities of investment analysts to monitor whether they are involved inconflicts of interest that undermine market integrity. However, in the pastthey have often focused on issues such as insider trading. Clearly, the recentcorporate scandals and legal actions against financial service providersindicate that a greater focus on conflicts of interest is needed in agencies thatsupervise securities markets.

The newly created PCAOB has the authority to monitor internal controls ataccounting firms, and the creation of this oversight board by Sarbanes-Oxleyis one of the most desirable features of this legislation. An important task ofthe PCAOB will be to ensure that auditors are independent of managementand report to audit committees. Also, the PCAOB will need to monitor andencourage best-practice compensation mechanisms inside accounting firmsthat continue to conduct auditing and management advisory services underthe same roof.

(v) Provide adequate resources to supervisors.Supervisors must havesufficient resources to monitor conflicts of interest. Supervision hasfailed when supervisors were starved for resources. In the 1980s,limited resources weakened the power of supervisors during the U.S.banking crises (e.g., see Federal Deposit Insurance Corporation 1997).Only after the recent emergence of serious conflicts of interest thatrocked the financial system did the SEC have its funding substantiallyraised. Starving supervisors of resources is often the result of stronglobbying efforts by the supervised industry. In the financial serviceindustry this problem may become worse during good times whenfinancial service providers are making huge profits. Althoughresources for supervisory oversight of the financial service industry

236 Mishkin

have risen recently, it is important that the lessons of the 1990s not beforgotten and that supervisors continue to be given adequate resourcesand their employees compensated to ensure high-quality expertise isavailable.

(vi) Enhance competitiveness in the rating agency industry.Whileanalysts, auditors, and most financial institutions operate in highlycompetitive markets, rating agencies are protected from competitionby high entry costs and the official sanction of their ratings byregulators. The barriers to competition for rating agencies need to bereduced to enhance the discipline of the market and ensure thatadequate resources are invested in their activities.

(vii) Prevent the co-option of information-producing agents by regulatorsand supervisors.Currently, a severe problem arises from the increasingstandardization of ratings and their designation for regulatorypurposes. This practice should be limited since it encourages firms topackage their financing to meet certain targets. Excessive dependenceof supervisors on rating agencies limits their effectiveness as monitorsand thus their potential contribution to information.

In a similar vein, overly standardized, detailed prescriptive accounting ruleshave the unintended consequence of decreasing the amount of informationin auditors’ reports. Instead, the focus should be on a “true and fair view” ofthe financial performance and financial position of the audited firm.

(viii) Avoid forced separation of financial service activities except inunusual circumstances.I am generally skeptical of forced separationof financial service functions to solve conflict of interest problems. Inmany cases, the market leads financial service firms to separateactivities, either with firewalls or by setting up separately capitalizedaffiliates, in order for the firms to attest to the quality of theinformation they provide and thus sell their services profitably. This isexactly what happened in the banking industry before the Glass-Steagall Act. In hindsight, we know that this act created a costly andrigid separation of commercial and investment banking that reducedthe efficiency of the financial system and prevented the development ofmarket mechanisms to contain conflicts.

Both the segregation of the audit business envisioned in the Sarbanes-OxleyAct and radical changes for analysts imposed by the global settlement by theNew York Attorney General, the SEC, and other regulators appear to bemisdirected and excessive responses to the collapse of the bull market.Because they segregate the activities of auditors and analysts, altering thecompensation and forcing a sharing of information by the latter, economiesof scope will be reduced and the quantity and quality of information maywell decline. Complete segregation is an extreme and usually inappropriateremedy. Litigation, industry standards, and supervisory oversight should be

Policy Remedies for Conflicts of Interest in the Financial System 237

sufficient to erect the limited firewalls needed in most cases, while themarket disciplines firms that are perceived to take advantage of conflicts ofinterest.

There is some role for regulations enforcing limited separation underunusual circumstances. For example, forcing banks to have separatelycapitalized affiliates to conduct investment banking, insurance, and othernon-banking activities makes good sense in order to limit extending thesafety net beyond banking activities. A government safety net for banks hasthe rationale that it is needed to prevent bank panics. However, a govern-ment safety net creates moral-hazard incentives for risk taking that requiremore extensive regulation and supervision to ensure the safety andsoundness of the banking industry. This problem is even more severebecause the government cannot credibly commit to avoid a too-big-to-faildoctrine. Extending the safety net to other financial service activities has amuch weaker rationale and would create further incentives for risk takingthat could be highly damaging.

(ix) Do not socialize information for the financial service industry.Socialization of information carries many hidden dangers for thequality of the information generated, and is generally unwarranted.Socialization could potentially take a variety of forms, includingofficial provision of certain services (e.g., research, auditing) and thefinancing of independent private sector services by taxation or a levy.I am, however, most skeptical of any remedy that mandates thesocialization of information production in financial markets. In itsextreme form, this approach negates the benefits of multiple,competing agents. Even where service providers themselves remain inthe private sector, there are threats to the quality of informationprovided. For example, if rating agencies are protected fromcompetition and their ratings are standardized and mandated for riskassessment, they have little incentive to devote effort to thoroughanalysis or to improve their assessment techniques. If auditors areinduced to produce opinions that are exclusively rule-based rather thanprinciple-based and the rules are tightly defined by the regulators, thenthey too become part of the regulatory system and do not contributeany independent judgment. A form of socialization has been incor-porated in the global settlement reached with the largest investmentbanks, where firms are required to purchase outside research and sharetheir own research. Although socialization of information productionwould reduce incentives to exploit conflicts of interest, it is likely toreduce the quality of information in the marketplace, and thereforemake the financial system less efficient, rather than more efficient.

238 Mishkin

These recommendations rely on the combination of market discipline, sup-plemented by mandatory disclosure of conflicts, and supervisory oversightto keep conflicts of interest from damaging information production in thefinancial system. In other words, policies should almost always be based onthe first three approaches to remedying conflicts of interest. Theseapproaches are complementary and are oriented to help make markets workbetter. Market discipline, supplemented by mandatory disclosure andsupervisory oversight is usually sufficient to control conflicts of interest. It isimportant to recognize that markets do not immediately create optimalstructures to solve conflict of interest problems. As the history of universalbanking suggests, financial markets move to manage conflicts effectivelyover time.

The bottom line is that radical solutions to conflict of interest problems thatinvolve socialization of information production or a very stringent separa-tion of financial service activities are likely to do far more harm than good.With increased disclosure of information and supervisory oversight, plusadditional reforms of rules governing audit opinions and official use andsanction of ratings, the problems created by conflicts of interest can beminimized. More radical approaches have the potential to reduce, ratherthan increase, the quality of information in financial markets, with the resultthat channelling funds to those with productive investment opportunities,which is so crucial to strong economic growth, could be severelycompromised.

References

Akerlof, G.A. 1970. “The Market for ‘Lemons’: Quality Uncertainty and theMarket Mechanism.”Quarterly Journal of Economics 84 (3): 488–500.

Ang, J.S. and T. Richardson. 1994. “The Underwriting Experienceof Commercial Bank Affiliates prior to the Glass-Steagall Act:A Reexamination of Evidence for Passage of the Act.”Journal ofBanking and Finance 18: (2): 351–95.

Bank for International Settlements. 2000. “Credit Ratings andComplementary Sources of Credit Quality Information.” BaselCommittee on Banking Supervision Working Paper No. 3. August.

Boot, A.W.A. 2000. “Relationship Banking: What Do We Know?”Journalof Financial Intermediation 9 (1): 7–25.

Brand, L. and R. Bahar. 1999. “Ratings Performance 1998.” Standard andPoor’s.

Calomiris, C.A.W. and C.M. Kahn. 1991. “The Role of Demandable Debt inStructuring Optimal Banking Arrangements.”American EconomicReview 81 (3): 497–513.

Policy Remedies for Conflicts of Interest in the Financial System 239

Chan Y.-S., S.I. Greenbaum, and A.V. Thakor. 1986. “InformationReusability, Competition and Bank Asset Quality.”Journal of Bankingand Finance 10 (2/3): 243–53.

Crockett, A., T. Harris, F.S. Mishkin, and E.N. White. 2004.Conflicts ofInterest in the Financial Services Industry: What Should We Do AboutThem?Geneva Reports on the World Economy 4 (International Center forMonetary and Banking Studies and Centre for Economic PolicyResearch: Geneva and London). CEPr.

Diamond, D.W. 1984. “Financial Intermediation and Delegated Monitoring.”Review of Economic Studies 51 (3): 393–414.

Edwards, F. 1979. Comment on “Implications of Growing Institutionalizationof the Stock Market.”Contemporary Studies of Economic and FinancialAnalysis,197–202. Greenwich, Conn.: JAI Press.

Federal Deposit Insurance Corporation. 1997.History of the Eighties—Lessons for the Future. Volume 1. Federal Deposit InsuranceCorporation.

Greenbaum, S.I. and A.V. Thakor. 1995.Contemporary FinancialIntermediation. Montréal: Dryden Press.

Keenan, S.C. 1999. “Historical Default Rates of Corporate Bond Issuers,1920–1998.” Moody’s Investors Services, Special Comment.

Kroszner, R.S. and R.G. Rajan. 1994. “Is the Glass-Steagall Act Justified?A Study of the U.S. Experience with Universal Banking Before 1933.”American Economic Review 84 (4): 810–32.

———. 1997. “Organization Structure and Credibility: Evidence fromCommercial Bank Securities Activities before the Glass-Steagall Act.”Journal of Monetary Economics 39 (3): 475–516.

Leland, H.E. and D.H. Pyle. 1977. “Informational Asymmetries, FinancialStructure, and Financial Intermediation.”Journal of Finance 32 (2):371–87.

Michaely, R. and K.L. Womack. 1999. “Conflict of Interest and theCredibility of Underwriter Analyst Recommendations.”Review ofFinancial Studies 12 (4) Special Issue: 653–86.

Parkash, M. and C.F. Venable. 1993. “Auditee Incentives for AuditorIndependence: The Case of Nonaudit Services.”The Accounting Review68 (1): 113–33.

Puri, M. 1996. “Commercial Banks in Investment Banking: Conflict ofInterest or Certification Role?”Journal of Financial Economics 40 (3):373–40.

240 Mishkin

Ramakrishnan, R.T.S. and A.V. Thakor. 1984. “Information Reliability and aTheory of Financial Intermediation.”Review of Economic Studies51 (3):415–32.

Santos, J.A.C. 1998. “Banking and Commerce: How Does the United StatesCompare to Other Countries?”Federal Reserve Bank of ClevelandEconomic Review 34 (4): 14–26.